Donald Henson, M.D.
GW Cancer Institute

Susan Cheng, PhD
Department of Computer Science

Computer Algorithms for Personalized Cancer Medicine

Medical progress is taking us toward a personalized form of medicine where proteogenomic expression profiles will enable physicians to predict a person¡¯s risk of disease, course of disease, and optimal therapeutic intervention. Prediction at the individual level will require computer algorithms that present outcome probabilities based on an analysis of complex medical data suitable for personalized medicine. We are proposing therefore to develop and test algorithms for a predictive system for application in personalized medicine. These algorithms will be able to incorporate any number and combination of predictive factors useful for physicians. The development of algorithms will follow recent techniques in data mining, a bottom up approach that depends on cluster analysis, and a top down approach based on the concept of group testing. Group testing entails continued stratification of predictor variables based on outcome.
   
 
  801 22nd ST, NW Suite T720B
Washington, DC 20052
Fax:202-994-4875 ibe@gwu.edu